Two questions sit behind most of what a communications team does: what just happened, and what does it mean. Today we're launching Truescope AI to answer both, faster than either has been answered before.
Truescope AI has two teammates, Assistant and Analyst. Both work through a conversational interface, and both are built into the reports and alerts you already use. Assistant takes the daily questions, grounded in your data. Analyst runs the deeper analysis that used to take days. Neither asks you to learn a new product; they live where your reporting already happens. The user stays in control of what reaches a stakeholder.
A faster read on the day, in one click
Assistant summarises your coverage. Because it sits on top of your queries, it knows your keywords and your data and where to focus. Log in, click once, and you have a digest of everything said across your queries that day.
It does three more jobs worth naming:
- It writes the summary at the top of a spike alert, so when volume, audience or engagement jumps on something sensitive, your brand or your chief executive in a run of negative coverage, you get four or five sentences on what's behind it.
- It turns plain English into Boolean strings, long the part of this work comms teams least enjoy.
- When you'd rather not hand the writing over completely, it drafts and you edit, so nothing reaches a stakeholder you haven't approved.
Research-grade reports in minutes, not weeks
Analyst is the deeper teammate. Media analysis has traditionally run on a monthly cycle. Analyst brings it down to minutes.
It works in three steps, the same three a good analyst has always worked through:
- It interprets your intent and asks clarifying questions until the brief is clear.
- It retrieves the data, item-level and aggregate, across sentiment, entities, key phrases, source authority, source bias and category, by volume, audience and engagement.
- It writes the report and builds the visualisations to match.
The retrieval and the charts are not drawn from a fixed set. Analyst generates what each question needs rather than picking from a menu, which is what lets it answer the question you actually asked.
If a question reaches past your usual coverage, Analyst can pull in real-time social data to fill the gap. No platform in this space, ours included, holds the entire social firehose, so it goes and gets what the brief needs at the time the report runs.
Built for the work comms teams already do
Truescope AI launches with more than 70 templates for the jobs comms teams run most, including:
- Share of voice
- Crisis assessment
- ESG reporting
- Influencer identification and profiling
- Auditing your owned social channels
Run one as it stands, duplicate and adapt it, or build your own with your team or ours.
Ask for share of voice and Analyst offers the template, confirms which competitors you mean, runs the five or six data jobs behind it, and lays the report out the way it should be laid out. Anything you can ask once, you can put on a schedule: a digest at nine each morning, a full analysis on the first of the month. And a spike alert is a starting point as much as a warning. Drop straight from it into a crisis assessment and get back a picture of the situation with recommendations attached.
There is also a layer of personalisation worth turning on. Tell the system which department you sit in, which organisation you work for and the areas you care about, and it reads your recent coverage to surface questions worth asking, one set for Assistant and one for Analyst. It also carries into anything you ask, so answers come back shaped for your role rather than written for everyone. None of it is required for the tools to run. It just sharpens what comes back.
Where we drew the line on AI
There is a lot of hype in the AI category, and comms teams are right to ask what happens when a model gets something wrong. So here is where we used it and where we didn't.
Relevance matching, enrichment, source authority and bias, classification: none of that is run by language models. It runs on systems we've hardened over four or five years, the same systems that scored your data before Truescope AI existed. The AI reads from that data. It doesn't generate it.
The rest is discipline:
- Low ‘temperature’ on every model call, for consistent output.
- Source grounding, so every claim cites the document behind it.
- Analyst has been in beta longer than anything else we've built, since last August, which gave us hundreds of pieces of feedback to act on.
- We stay on the frontier models as they get better.
It would be wrong to claim it never errs. Inside your own coverage it's very accurate. Send it across the global data pool and it can lose context, the classic case being Apple the company against the fruit, and a summary can underweight an item you would have led with. We don't pretend that's solved. We make it easy to audit instead: Analyst surfaces every item behind a report, and a fix is usually one prompt away.
What's next
What we're launching is good and useful. It isn't the destination. This category is going to keep changing quickly, probably faster than anything I've seen in this industry, and the six years we spent building the data foundation are what let us keep pace. The interface will change. The question won't.
If you'd like to see Assistant and Analyst run on your own coverage, book a demo and we'll walk you through it.
See how you can get more signals and insights from your media intelligence platform
See how you can get more signals and insights from your media intelligence platform
